Myo-inositol is a naturally-occurring known substance represented by the following steric structural formula (A).

Scyllo-inosose is a known substance represented by the steric structural formula (B).

Furthermore, scyllo-inositol is a known substance represented by the following steric structural formula (C).

Scyllo-inositol is one of stereoisomers of myo-inositol and is a substance widely found among animals and plants. Scyllo-inosose is a compound having a structure in which an axial hydroxyl group at the second position of myo-inositol is oxidized, and exists generally as a natural compound.
Scyllo-inositol is a substance expected for applications such as a therapeutic agent for an Alzheimer disease (see Non-Patent Document 1), a raw material for synthesis of bioactive substances (Patent Document 1), or a raw material for synthesis of liquid crystal compounds (Patent Document 2).
Examples of a method of producing scyllo-inosose or scyllo-inositol by means of a chemical synthetic procedure include: (i) a method of obtaining scyllo-inositol by reducing hexahydroxybenzene with Raney nickel (Non-Patent Document 2); (ii) a method of obtaining scyllo-inositol by reducing scyllo-inosose obtained from a glucofuranose derivative through a reaction involving five steps (Non-Patent Document 3); (iii) a method of obtaining scyllo-inositol using as a raw material cis-trioxa-tris-homobenzene through a reaction involving four steps or more (Non-Patent Document 4); and (iv) a method of obtaining scyllo-inositol including oxidizing myo-inositol with a platinum catalyst to thereby obtain scyllo-inosose, and subjecting the scyllo-inosose to esterification followed by reduction and hydrolysis (see Patent Document 2).
As a method of converting myo-inositol into scyllo-inositol using a microorganism, a method using a bacterium belonging to the genus Agrobacterium is known (Patent Document 3). However, this method is not applicable for an industrial-scale production because of low yield of scyllo-inositol and generation of other converted products.
Meanwhile, a bacterium belonging to the genus Acetobacter (see Non-Patent Document 5) is known to act on myo-inositol to absorb oxygen to thereby oxidize myo-inositol into scyllo-inosose. However, its detailed mechanism has not been studied.
The enzyme which oxidizes myo-inositol into scyllo-inosose (myo-inositol 2-dehydrogenase) has been reported from a number of organisms such as animals, algae, yeasts, and bacteria, and it is an enzyme that widely exists in nature. Examples of a typical microorganism having the enzyme include Aerobacter aerogenes (see Non-Patent Document 6), bacteria belonging to the genus Bacillus (Non-Patent Document 7 and 8; Patent Documents 4-6), and bacteria belonging to the genus Pseudomonas (Non-Patent Document 9 and 10).
However, the myo-inositol 2-dehydrogenases in those reports are NAD+-dependent enzymes, therefore they require NAD+ or NADP+ for oxidation. When the enzyme is subjected to an industrial-scale reaction, fermentative production must be employed in order to recycle those co-enzymes, resulting in decomposition of part of substrates. In addition, there had been problems in industrial-scale production such that the concentration of the substrate should be kept low.
Meanwhile, there is a report of the presence of a scyllo-inositol dehydrogenase in a bovine brain and a fat tissue of a cockroach (Non-Patent Document 11). When scyllo-inosose as a substrate is reduced by this enzyme with NADPH, both of scyllo-inositol and myo-inositol are reported to be generated. However, the enzyme has low substrate specificity, a highly purified enzyme was not used, and other properties are unknown, therefore the enzyme may be an alcohol dehydrogenase having low substrate specificity. Therefore, the enzyme has not been described in Handbook of Enzymes (published by Asakura Shoten). As described above, although reports on animals exist, it has not been ascertained whether these reports are true.
Furthermore, there is also a known method of producing scyllo-inositol by chemically reducing scyllo-inosose produced by microbial oxidation (Patent Document 7). Since the substance obtained by the chemical reduction of scyllo-inosose is a mixture of scyllo-inositol and myo-inositol, the mixture had to be desalted and purified, followed by separation of scyllo-inositol having low solubility from the concentrated solution by crystallization. Thus, those methods have required many operations and thus there has been a room for improvement with respect to the yield of scyllo-inositol. Under such circumstances, the development of a method of producing purified scyllo-inositol from a mixture of scyllo-inositol and myo-inositol which is obtained by reduction of scyllo-inosose, or the like, has been expected in order to produce scyllo-inositol conveniently and efficiently.
When scyllo-inosose is reduced using NaBH4 in a solution, the solution after the reaction contains myo-inositol, scyllo-inositol, and a small amount of a scyllo-inositol/boric acid complex. For such scyllo-inositol/boric acid complex, there has been known a method of obtaining scyllo-inositol involving: filtrating the complex as a precipitate; dissolving the precipitate in diluted sulfuric acid; adding thereto methanol to subject it to azeotropy with boric acid; removing the boric acid; and desalting the remaining solution using an ion exchange resin (Non-Patent Document 12).
The scyllo-inositol/boric acid complex is a substance represented by the following steric structural formula (D).

However, in the above-described method of reducing scyllo-inosose using NaBH4, the ratio of the generated scyllo-inositol/boric acid complex is low, and scyllo-inositol is also generated in the solution. Therefore, the complex and components in the solution had to be separated to thereby obtain scyllo-inositol from each of those. Furthermore, a large amount of an organic solvent has been required to obtain scyllo-inositol from the complex, there has been a room for improvement in an economical viewpoint. Thus, there have been demanded a method of producing scyllo-inositol conveniently and efficiently in industrial-scale production.
[Patent Document 1] U.S. Pat. No. 5,412,080
[Patent Document 2] DE 3,405,663
[Patent Document 3] JP09-140388A
[Patent Document 4] JP04-126075A
[Patent Document 5] JP05-192163A
[Patent Document 6] JP06-007158A
[Patent Document 7] JP2003-102492A
[Non-Patent Document 1] The Journal of Biological Chemistry (US), 2000, vol. 275, No. 24, p. 18495-18502
[Non-Patent Document 2] Journal of the American Chemical Society (US), 1948, vol. 70, p. 2931-2935)
[Non-Patent Document 3] Journal of the American Chemical Society (US), 1968, vol. 90, p. 3289-3290
[Non-Patent Document 4] Angewandte Chemie (Germany), 1973, vol. 85 p. 1110-1111
[Non-Patent Document 5] The Journal of Biological Chemistry (US), 1948, vol. 174, p. 173-188
[Non-Patent Document 6] Archives of Biochemistry and Biophysics (US), 1956, J, Larner et al., vol. 60, p. 352-363
[Non-Patent Document 7] The Journal of Biological Chemistry (US), 1979, vol. 254, p. 7684-7690
[Non-Patent Document 8] Microbiology (US), 1994, vol. 140, p. 2289-2298
[Non-Patent Document 9] Monatshefte fur Chemie (Germany), 1969, vol. 100, p. 1327-1337
[Non-Patent Document 10] Journal of Bacteriology (US), 1977, vol. 131, p. 872-875
[Non-Patent Document 11] Biochemical and Biophysical Research Communications (US), vol. 68, p. 1133, 1976
[Non-Patent Document 12] Journal of Organic Chemistry (US), 1958, vol. 23, p. 329-330